Abstract
Aspiration detected in the fiberoptic endoscopy evaluation of swallowing (FEES) has been inconsistently associated with pneumonia, with no evidence of the risk of pneumonia from other alterations in swallowing safety detected in FEES. We conducted a dynamic, ambidirectional cohort study involving 148 subjects at risk of dysphagia in a tertiary university hospital. Our aim was to determine the risk of pneumonia attributed to alterations in swallowing safety detected during FEES. We used multivariate negative binomial regression models to adjust for potential confounders. The incidence density rate (IR) of pneumonia in patients with tracheal aspiration of any consistency was 26.6/100 people-years (RR 7.25; 95% CI: 3.50-14.98; P < 0.001). The IR was 19.7/100 people-years (RR 7.85; 95% CI: 3.34-18.47; P < 0.001) in those with laryngeal penetration of any consistency and 18.1/100 people-years (RR 6.24; 95% CI: 2.58-15.09; P < 0.001) in those with pharyngeal residue of any consistency. When adjusted for aspiration, the association of residue and penetration with pneumonia disappeared, suggesting that their risk of pneumonia is dependent on the presence of aspiration and that only aspiration is independently associated with pneumonia. This increased risk of pneumonia was significant in uni- and multivariate negative binomial regression models. We found an independently increased risk of pneumonia among patients with dysphagia and aspiration detected during FEES. Alterations in the oral and pharyngeal phases of swallowing, without aspiration, did not increase the risk of pneumonia.